> For the complete documentation index, see [llms.txt](https://docs.pharos.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.pharos.ai/configuration-guide/integrations/ai-agents-mcp/skill-catalog.md).

# Skill Catalog

The full catalog of Pharos operate skills for MCP-connected AI agents — front-door router, reference skills, debugging workflows, and an alerting on-ramp.

The Pharos operate skills are agent-agnostic Markdown playbooks that run on top of the `@pharosai/mcp` server. This page lists every skill and shows how they hand off to one another. For prerequisites, permission tiers, and how to invoke skills in your agent, see [AI Agents & MCP](https://docs.pharos.ai/configuration-guide/integrations/ai-agents-mcp.md).

***

## Front Door

Start here whenever the user hasn't already named a destination.

| Skill                | Purpose                                                                                                                                            |
| -------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------- |
| **`operate-pharos`** | Turns a plain-language question, at any level of specificity, into the right diagnostic or action workflow. The default entry point for a session. |

***

## Reference Skills

These are loaded automatically by other skills as context—you never invoke them directly.

| Skill                      | Purpose                                                                                                                                           |
| -------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------- |
| **`understanding-pharos`** | How Pharos stores, groups, and exposes diagnostic data: Logs vs. Issues, org orientation, log provenance, field semantics, and query patterns.    |
| **`understanding-triton`** | Deep dive on Triton-authored logs: Area/Type/Level anatomy, data richness by level, transaction stitching, and when to bridge to instrumentation. |

***

## Debugging Workflows

The read/write loop, from finding a problem through acting on it.

| Skill                    | Tier         | Purpose                                                                                                                                         |
| ------------------------ | ------------ | ----------------------------------------------------------------------------------------------------------------------------------------------- |
| **`report-org-health`**  | Read         | Sweep the org for what's broken now—rank issues by impact, segment by area and recency, optionally correlate spikes with recent config changes. |
| **`investigate-logs`**   | Read         | Start from a raw symptom or spike with no Issue identified yet—scope a query, triage the sample by parent Issue, route onward.                  |
| **`investigate-issue`**  | Read         | Diagnose a specific, already-identified Issue—stats, existing analysis, correlated logs, and a root-cause read.                                 |
| **`analyze-root-cause`** | Read         | Launch or retrieve Pharos's AI-generated root-cause analysis for a known Issue, interpreted against the logs rather than taken at face value.   |
| **`capture-a-trace`**    | Read + Write | Set up go-forward log capture for a symptom that isn't currently landing in Pharos, then harvest and clean up what the trace collects.          |
| **`triage-and-promote`** | Read + Write | Act on a diagnosed Issue—update status, assignee, or priority, and optionally promote it to a Jira or Azure DevOps ticket.                      |

***

## Alerting On-Ramp

| Skill                 | Tier  | Purpose                                                                                                                                                                                                                  |
| --------------------- | ----- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| **`set-up-alerting`** | Admin | A basic observability on-ramp: create, list, and toggle notification and basic threshold alerts. More involved setup—retention, masking, digests, limits—is named and routed to the Pharos admin UI with a docs pointer. |

{% hint style="info" %}
**`set-up-alerting` is an on-ramp, not a rule editor.** It authors the common alert types; anything more involved is routed to the admin UI. See [Notifications](https://docs.pharos.ai/configuration-guide/notifications.md) and [Log Retention](https://docs.pharos.ai/configuration-guide/log-retention.md) for full rule configuration.
{% endhint %}

***

## How Skills Hand Off

Skills suggest the natural next step rather than leaving you at a dead end. Two common paths:

* **Diagnose then act:** `investigate-issue` → `analyze-root-cause` → `triage-and-promote`
* **Broad to specific:** `report-org-health` → `investigate-issue` on a flagged item

When a workflow finds an area that is dark—nothing logging at all—it recommends the appropriate [Triton instrumentation](https://docs.pharos.ai/platform-overview/pharos-platform-architecture/logs/log-ingestion-methods/pharos-triton-instrumentation.md) skill (`instrument-apex`, `instrument-lwc`, `instrument-flows`) as a bridge to start capturing data.

***

## Related

* 🤖 [AI Agents & MCP](https://docs.pharos.ai/configuration-guide/integrations/ai-agents-mcp.md) — prerequisites, permission tiers, and how to connect your agent
* 🛠️ [Pharos Triton Instrumentation](https://docs.pharos.ai/platform-overview/pharos-platform-architecture/logs/log-ingestion-methods/pharos-triton-instrumentation.md) — the write side that produces the logs these skills read
